This article was downloaded by: [University of Nevada Las Vegas] On: 03 September 2011, At: 19:57 Publisher: Routledge Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Studies in Conflict & Terrorism Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/uter20 Failed States and the Spread of Terrorism in Sub-Saharan Africa Tiffiany Howard a a Department of Political Science, University of Nevada, Las Vegas, Las Vegas, NV, USA Available online: 23 Oct 2010

To cite this article: Tiffiany Howard (2010): Failed States and the Spread of Terrorism in Sub-Saharan Africa, Studies in Conflict & Terrorism, 33:11, 960-988 To link to this article: http://dx.doi.org/10.1080/1057610X.2010.514696

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan, sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. Studies in Conflict & Terrorism, 33:960–988, 2010 Copyright © Taylor & Francis Group, LLC ISSN: 1057-610X print / 1521-0731 online DOI: 10.1080/1057610X.2010.514696

Failed States and the Spread of Terrorism in Sub-Saharan Africa

TIFFIANY HOWARD Department of Political Science University of Nevada, Las Vegas Las Vegas, NV, USA

Plagued by systematic state failure, sub-Saharan Africa’s failed states have helped facilitate internationally sponsored terrorist networks and operations. However, until recently, this type of activity was primarily relegated to North Africa and the Horn. But that has begun to change. Now, what was once a seemingly benign terrorist presence in sub-Saharan Africa is starting to transform into a movement, with states such as Nigeria, Sierra Leone, Liberia, and the Democratic Republic of the Congo (DRC) now lending arms, financial support, and radical militants to the extremist jihadist movement of internationally sponsored terrorist organizations such as Al Qaeda. Further, with the recent bombings in Kenya and Somalia, and the failed bombing attempt on a U.S. airliner by a Nigerian national, it is becoming increasingly evident that internationally sponsored terrorist networks have found a permanent home in sub-Saharan Africa and within the hearts and minds of its people, which poses significant challenges for the international community, given the region’s patchwork of failed states, where terrorists can easily hide and thrive. Consequently, this study discusses how the conditions of state failure have fostered support for internationally sponsored terrorism in sub-Saharan Africa. Terrorist groups are now actively recruiting more militants from within the region and popular support for extremist acts is on the rise in sub-Saharan Africa. Thus, the article argues sub-Saharan Africa will soon become the site for the next generation of terrorists, and the next wave of terrorist activity.

With the Global War on Terrorism (GWOT), Africa has been dragged into the middle of the discussion, despite that the majority of conflicts and terrorist activity in the region are not connected to international sponsorship or a comprehensive agenda against a specific target in the West. The Lords Resistance Army in Uganda, the insurgents in the Niger Delta, the Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 extremists in Kenya and Nigeria, the militias in the Democratic Republic of the Congo, and a collection of rebel groups and insurgent movements throughout the region are the principal security threats to the citizens of Africa—not Al Qaeda, the Taliban, or any other internationally recognized terrorist organization on the U.S. State Department’s list. Yet, Africa has become a target in the GWOT, but Africa’s greatest offense has been that its network of failed and weak states are simply unable to sufficiently monitor its borders or govern its territory, thus opening the door for internationally sponsored terrorist networks

Received 7 December 2009; accepted 14 February 2010. Address correspondence to Tiffiany Howard, Assistant Professor, Department of Political Sci- ence, University of Nevada, Las Vegas, 4505 S. Maryland Pkwy., WRI B227 Box 455029, Las Vegas, NV 89154, USA. E-mail: tiffi[email protected]

960 Failed States and Terrorism in Sub-Saharan Africa 961

to take up residence within the region, where they have gone on to plan attacks against American, European, and Israeli targets, built local cells with the capacity to attack in Africa or recruit for operations elsewhere, and found sympathetic agents within the population to provide safe haven and financial support for terrorists and their operations (Harbeson and Rothchild 2008). Africa is undoubtedly plagued by systematic state failure in that the region lacks strong governance, comprehensive economic development, and fails to provide security to its citizens and order in its territories. As a result, Africa’s failed states have helped facilitate internationally sponsored terrorist networks and operations. However, until re- cently, this type of activity was primarily relegated to North Africa and the Horn of Africa. This is because North Africa and the Horn include states collectively known as the Maghreb, which share historical, cultural, and religious ties with the Middle East. Consequently, North Africa and the Horn are far more integrated into the network of international extremist Islamic terrorist activity than sub-Saharan Africa. For example, Algerians and Moroccans have consistently been involved in many attacks in Europe and comprise the majority of Africans who have reportedly traveled to join the jihadist movement in Iraq. Further, the development of the radical Islamic Courts Movement in Somalia has reinforced the economic, political, religious, and ethnic linkages between the Horn of Africa and the Middle East, as people, arms, money, and materials move from the Arabian Peninsula, across the Red Sea, through Somalia and Sudan into the heart of Africa. What was once simply a seemingly benign terrorist presence in sub-Saharan Africa is starting to transform into a movement, with states such as Nigeria, Sierra Leone, Liberia, and the Democratic Republic of the Congo (DRC) now lending arms, finan- cial support, and radical militants to the extremist jihadist movement of internationally sponsored terrorist organizations such as Al Qaeda. Furthermore, with the recent bombings in Kenya and Somalia, and the failed bombing attempt on a U.S. airliner by a Nige- rian national, it is becoming increasingly evident that internationally sponsored terrorist networks have found a permanent home in sub-Saharan Africa, and within the hearts and minds of its people, which poses significant challenges for the international commu- nity, given the region’s patchwork of failed states, where terrorists can easily hide and thrive. Thus, while internationally sponsored terrorist organizations have been present in sub-Saharan Africa for the past forty years, the region has seemingly been immune to terrorist attacks, but that is changing. With successful independence movements, and the emergence of democratic states, the people of sub-Saharan Africa were hopeful for the region’s future; but with constant civil wars, brutal dictatorships, and violent struggles for power, sub-Saharan Africa has become a region of failed states. As a consequence of Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 sustained failure in the region, the people of sub-Saharan Africa have become disillusioned with its leadership and political institutions, thus making them more susceptible to the radical ideologies of internationally sponsored terrorist groups that promise economic prosperity and political power; and it is this phenomenon that represents the crux of this analysis. This study discusses how the conditions of state failure have fostered public support for internationally sponsored terrorism in sub-Saharan Africa. Terrorist groups are now actively recruiting more militants from within the region and popular support for extremist acts is on the rise in sub-Saharan Africa. Thus, this article argues that sub-Saharan Africa will soon become the site for the next generation of terrorists, and the next wave of terrorist activity. 962 T. Howard

General Proposition and Major Arguments Unlike other regions, such as the Middle East and South Asia, where extensive research has been conducted on the conditions that foster popular support for terrorism, there is a glaring absence of research on this phenomenon and its presence in sub-Saharan Africa. This is largely because acts of terrorism are not as prevalent in this region as they are in the rest of the world, but in recent years that has begun to change with the major incidents in Mombasa, Kenya and Mogadishu, Somalia (Mohammed 2009; Snowdon and Johnson 2007). This study evaluates the relationship between state failure in sub-Saharan Africa and the growing presence of terrorist and insurgent groups in these states, along with growing support for the use of political violence, in an effort to illustrate the very real likelihood that Africa’s failed states will soon become the breeding ground for extremist ideologies and a new generation of terrorists. While it is tempting to deduce the argument to the simple explanation that weak and failed states would naturally be attractive to terrorist groups and insurgents, given the absence of a functioning security force, it is asserted here that the relationship is more nuanced than this. There is certainly the element of a pull factor present—in that actors engaging in political violence are better able to carry out operations in failed states; and up until now, the majority of studies have remained solely focused on this element—the attractiveness of failed states to terrorists and insurgents (Newman 2007; Freedman 2002; Linden 2002; Menkhaus 2002). However, the article raises another argument in this dynamic and that is the possibility that citizens of failed states are attracted to political violence because of the deteriorating conditions within this type of state. A critical element that has remained absent from this discussion is the decision-making process of ordinary citizens to engage in terrorist and insurgent activity because the state can no longer fulfill its responsibilities to its citizens. Individuals living in failed states are attracted to political violence because the system is broken—the state has failed in its duty. Rather than evaluate the patterns of political violence in relation to Africa’s weak states using event data, the goal is to understand individuals’ perspectives regarding the use of violence as a political tool when the state has failed. Using barometer survey data, it is found that there is an insidious pattern of deprivation within failed states that drives ordinary citizens to engage in and support political violence. This study helps further scholarly discourse by suggesting an alternative root cause of this type of violence that has been overlooked by global indicator models of terrorism, because up until now they have ignored individual level data, thus they have also ignored the basic tenets of human psychology and what drives seemingly rational human beings to commit irrational acts of violence. Failed states threaten an individual’s survival, which ultimately drives them to obtain Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 tangible political and economic resources through other means, which include the use of political violence. Further, the major implications for the international community are that this pattern of deprivation makes individuals in these states more susceptible to the influence of internationally sponsored terrorist groups. As a consequence, failed states are breeding grounds for terrorists, who then export their radical ideologies to other parts of the world to create terrorist threats across the globe. Thus, the article argues that the GWOT will remain a futile endeavor if the international community ignores the importance of comprehensive state-building that incorporates the sustained development of strong political and economic institutions within developing societies (Howard 2008, 2010). The importance of this research is twofold. First, this study joins the existing body of research on weak states, fragile states, and failed states (Rotberg 2003, 2004; Esty et al. Failed States and Terrorism in Sub-Saharan Africa 963

1998; Goldstone et al. 2000, 1999; Miliken 2003; Howard 2008, 2010; Debiel and Klein 2002; Chesterman, Ignatieff, and Thakur 2005; Chandler 2009; Zartman 1995) in that it outlines a discrete set of indicators of state decline. However, it goes beyond these works by evaluating how these factors work together to weaken a state and create chaos and anarchy within a nation where terrorism and violence then become a legitimate avenue to obtain political and economic resources. Beyond establishing a typology of state failure, this study is also particularly timely, and unique from other works on terrorism in its approach to understanding alternative root causes of this type of violence, beyond the typical factors given, such as economic decline, religious radicalism, and levels of frustration (Martin 1987; Huntington 1991; Sorli, Gleditsch, and Strand 2005; Piazza 2006, 2007; Newman 2007; Tikusis 2009). It is argued that, unlike previous works, which focus on macro-level terrorist incidents and specific country cases that paint a broad picture of how weak states contribute to political violence, this study probes deeper and seeks to understand what it is about the process of state failure that influences an individual’s decision to support and commit an act of political violence.

Africa’s Failed States This study restricts the unit of analysis to sub-Saharan Africa, because this region represents a unique case when it comes to the discussion of state failure and political violence. State failure is an apparent pandemic in the region of sub-Saharan Africa. With the exception of Western Sahara and Swaziland, fifty-two of Africa’s fifty-four countries are on the 2009 Failed State Index as states at the critical levels of warning or alert. The author believes that state failure remains such a problem in sub-Saharan Africa because the region was never organized along the cotemporary state model that is recognized today. As a result, sub-Saharan Africa must now struggle to build political and economic institutions capable of sustaining a functioning state. Research dedicated to understanding post-colonial African states acknowledges that many political scientists try to impose a concept of statehood based on the models of Weber that are not appropriate given the context of ethnic and tribal identities unique to post- colonial African nations (Jackson and Rosberg 1982; Bates 1981; Widner 1995; Herbst 1996). Africa was never comprised of classically defined sovereign states that maintained a monopoly over the territory within its boundaries. Furthermore, many colonial powers that established a presence in Africa made no effort to extend the administrative authority of the government beyond the capital city. “In most cases, the colonial governments were little more than elementary bureaucracies, with limited personnel and finances, and were more comparable to rural country fiefdoms in Europe than modern independent states” (Jackson Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 and James 1993, 139). After achieving independence, several African nations continued this trend and did little to extend the administrative authority of the government beyond the urban centers. Consequently, considering Africa’s historical background, Herbst (1996) argues that the failure to resuscitate failed states or prevent weak states from failing in Africa is the result of the international system’s legitimization and recognition of several African nations that were really never states to begin with. The vast majority of reports and analytic studies that examine the state continue to use existing nation-states as their unit of analysis. Yet borders manufactured by colonial powers make the concept of statehood in Africa very difficult to achieve. Given that post-colonial Africa lacks a history of state development, it is very difficult for these contemporary states to create strong states. Many post-colonial African 964 T. Howard

nations are experiencing episodes of state failure because they simply lack a blueprint. Thus, it is difficult for states that lack a history of prior state development to establish strong and effective states today because the political and economic institutions necessary to promote state building were never cultivated in the past.

Terrorism, Insurgency, and Political Violence in Africa Political violence in Africa is more widespread and commonplace than many may realize. Africa’s most tragic failed states—Sudan, Sierra Leone, and Somalia—are all well-known sponsors of terrorism, and at various time periods in their histories have all provided safe haven for dangerous terrorist networks, such as Al Qaeda. As if managing terrorism in sub-Saharan Africa was not difficult enough, this region must also contend with insurgent activity. Insurgent movements in sub-Saharan Africa have waged a campaign of terror against citizens within their own country, which could rival the attacks of any international terrorist operation. For example, while some would classify Sierra Leone’s Revolutionary United Front (RUF) as primarily an opposition movement, the U.S. State Department has named the RUF as a terrorist group, pointing to the violence that characterized their campaign of terror against the weak government during the nineties. Regardless of one’s position on the appropriate classification of the RUF, the RUF’s re- lationship with Al Qaeda through the diamond smuggling operation, and its harboring of members of this terrorist organization within Sierra Leone is well known. Insurgent and terrorist activity is also well documented in Sudan. Backed by the Sudanese government, in 2003 the paramilitary organization the Janjaweed began imple- menting a brutal campaign against the popular resistance movement, the Sudan Liberation Movement/Army (SLM/A), which is comprised of three non-Arab groups—the Masalit, Fur, and Zaghawa. The conflict between the two forces plunged the Darfur region into anarchy, and led to the internal displacement of over two million forced migrants, and more than 200,000 refugees fled to neighboring Chad. At the same time, Sudan remains on the U.S. State Department’s list of states that sponsor terrorism, given that the state has provided safe haven for Hezbollah, the Palestinian Islamic Jihad, Hamas, the Egyptian Islamic Group, and the Abu Nidal organization, all well-known and very dangerous terrorist organizations (Linden 2002). With the enduring presence of political violence in sub-Saharan Africa, this region represents an ideal case for analysis. Citizens have likely been affected by political violence in some capacity, and therefore will certainly have definitive opinions regarding the use of political violence as a tool for political gains. However, before turning to a discussion of the hypotheses, there remains one facet of the argument that requires elaboration, and that is determining what is political violence, and the activities that comprise this broad term. Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011

Defining Political Violence—Rebel Groups, Insurgent Activity, and Terrorism There exist strong divisions within the literature regarding the definition of terrorism. This is mainly due to regional and cultural differences that manifest themselves within the research and literature. What one scholar defines as terrorism, the other considers a resistance movement. This polarization within the literature can be attributed to the notion that the word “terrorism” in itself is fraught with a myriad of racial, cultural, and religious complexities. It is that duality of perception that makes the process of arriving at a unified definition extremely challenging. Failed States and Terrorism in Sub-Saharan Africa 965

The usage of the term terrorism can be traced back to the French Revolution, and the contemporary Oxford English Dictionary illustrates this within the definition of terrorism as it references the French Revolution as an example:

Terrorism: A system of terror. 1. Government by intimidation As directed and carried out by the party in power in France During the revolution of 1789–1794. (Hoffman 2006, 2)

At the time of the French Revolution, terrorism was viewed as a positive action to illustrate dissatisfaction with the government. The widespread view was that the resistance of the people against the government was legitimate in light of the atrocities the government committed against the people. The same ideals were adopted during the American Revolu- tion, which the French went on to borrow. Thus, because of the nature of the violence that took place during the French Revolution, the French conception of terrorism has come to represent the basic tenant of terrorism—violence against the government. Yet violence against a regime in and of itself does not represent terrorism. The tension that exists regarding what should be defined as terrorism, and what should be considered a legitimate resistance movement or revolution, has plagued scholars for many years, and there still remain sharp divisions within the literature (Chomsky 1988; Craig 1988; Long 1990; Rapoport 1999; Zanini 1999; Cooley 2000; Wilkinson 2000; Cillufo 2001; Addison 2002; Franks 2006; Hoffman 2006). While much of the literature on political violence has focused on terrorism, it is important to point out that there are two strategic types of terror; the singular difference is the ideology of the acting group. Insurgent groups1 often organize within ungoverned territory, just as terrorist groups do; however, the goals of insurgents are vastly different from the goals of terrorists, despite the conflation of these two groups. Insurgency is defined as “an organized movement aimed at the overthrow of a constitutional government through the use of subversion and armed conflict” (Rabasa et al. 2007, 3). Although, insurgents rely on terror as a coercive tool, their overall goal is to conduct military operations, through the use of guerilla tactics, or some other strategy for the sole purpose of seizing and acquiring territory where they can create a “liberated zone” or “counter-state” (Rabasa et al. 2007). Insurgents also tend to rely on the aid and support of the mass population, from which they recruit members, enjoy popular support, and provide material resources as compensation. However, terrorists are often isolated, and they operate in small, clandestine networks. Although their aims may benefit the populace, they do not tend to insert themselves within the mass population as insurgents do. Terrorists are often characterized by their attacks, which are generally targeted at non-combatants, hence the widespread outrage and reaction Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 to large-scale terrorists attacks. Finally, the aims of terrorists are sometimes viewed as “millenarian, unachievable,” and idealistic (Rabasa et al. 2007, 3). Given the presence of political violence in Africa, this analysis is focused specifically on determining the impact the economic and political conditions in the region have on an individual’s support for the use of political violence as a legitimate political tool. The Afrobarometer Survey allows this study to do this in that it relies on a measure of political violence, where it asks respondents to indicate whether or not they sometimes consider it necessary to use violence in support of a just cause. The only limitation of the survey on this issue is that no distinction is made between the variant types of political violence that typically come in the form of rebel or insurgent activity, and terrorism. The conflation of all types of political violence into a single measure raises the issue of whether or not 966 T. Howard

individuals actually make the distinction between these types of political violence. It is here argued that while there are unique motivators for each type of violent act, insurgent activity and terrorist attacks are often times indistinguishable. Therefore, while it is acknowledged that insurgent activity and terrorist activity are discrete types of political violence, it is also argued that these are simply labels and that individuals often do not make a distinction between them. Terrorists and insurgents are organized around different agendas and maintain varied objectives, and yet, their actions, and use of terror as a strategy, are nearly identical. Consequently, the ideological and fundamental differences between insurgents and terrorists are acknowledged, but at the same time, this analysis does not distinguish between the two groups. Therefore, adopting the definition set forth by Hoffman in Inside Terrorism (2006), this article defines acts of terror as “[an] abhorrent act of violence perceived as directed against society...” (2006, p.1). Consequently, whether or not these groups are classified as insurgents, rebel groups, drug cartels, paramilitary organizations, revolutionary movements, or ultimately as terrorists, the article considers any strategic activity with the intent of inducing fear, that is based on coercion, and for the purpose of threatening harm with the goal of forcing political or economic concessions to be an act of political violence.

The Hypotheses The main hypothesis that guides this analysis is that failing and failed states are breeding grounds for political violence. This is mainly due to the weak rule of law in these fragile states, and the complete absence of a security infrastructure capable of monitoring territorial borders. However, it is also because the conditions in failing and failed states have the potential to drive individuals to resort to violent activities in order to achieve some measure of security, and obtain tangible political and economic goods.

State Failure Strong states provide one basic, fundamental public good, which weak states lack: the public good of security (Rotberg 2003). Without security, it is difficult for states to provide other basic goods that indicate a state is stable. Therefore, the article sets forth five hypotheses that capture the dimensions of state weakness.

The Public Good of Security H1: Individuals who feel the state does not provide adequate personal security Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 are likely to hold support for the use of political violence.

This hypothesis is grounded in the notion that individuals who have been victims of a violent act will turn to political violence as a means to secure some measure of security. This is because the state has failed to provide security. Consequently, these individuals are more likely to engage in violent acts against the state, or support groups that may commit violent acts, but at the same time provide ad hoc security. This particular phenomenon is based on what some sociologists have identified as the key factor that contributes to gang membership. Studies have found that adolescent males who have witnessed violence, or have been a victim of violence, no longer believe the state and its security personnel are capable of protecting them (Adamson 1998; Akers 1998; Akers and Gang 1999; Elliot Failed States and Terrorism in Sub-Saharan Africa 967

and Menard 1996; Flannery and Huff 1999; Klein and Maxson 2006: Seifert 2006). As a consequence, they seek protection from alternative groups, such as a gang, to provide this protection, even as the gang goes on to commit violent acts. Seifert applies this particular cognitive behavior to not only the recruitment of members into gangs, but cults and terrorist organizations as well (2006). She argues that adolescent exposure to war and genocide leads to an increased likelihood that these individuals will grow up to become violent and participate in social violence, such as a terrorist organizations (2006). In sub- Saharan Africa, where both adult and adolescent exposure to violence is very common, it is asserted here that there is the likely possibility that citizens in this region are more accepting of political violence than in regions where exposure to violence is less prevalent. Consequently, it is argued that personal insecurity has the potential to lead to support for and participation in political violence.

The Presence of the State H2: The greater the presence of the state, the less likely an individual will participate in and hold support for political violence.

In every known case of state failure, the absence of the state’s influence is captured by the existence of ungoverned territory. Ungoverned territories are characterized by large stretches of land within a state that are without rule of law exercised by the central govern- ment. These spaces are typically located in rural areas, beyond the influence of the central government, or in mountainous locations where there is rough terrain (Rabasa et al., 2007; Howard 2010). The existence of ungoverned territory within a weak state is a recipe for dis- aster. Without the presence of the government and security personnel, non-state actors such as rebel groups, terrorist cells, paramilitary units, and insurgents can organize themselves in these spaces and engage in illegal and dangerous activities without having to contend with the presence of the central government. Given what is known about ungoverned territory, this study hypothesizes that the greater the presence of the state, the willingness of an individual to engage in political violence, or support the use of political violence is less likely.

State Authority H3: The greater the authority and legitimacy of the state, the less likely an individual will participate in and hold support for political violence.

The logic behind this hypothesis is that the leadership of a state sets the tone for the state. If the state authority is perceived as corrupt, or as having seized power through illegitimate means, it becomes difficult for the state to convince citizens to adhere to basic laws and Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 institutional rules (Levi 1988). The only exception is when state leaders use violence and coercion to maintain authority and legitimacy, in lieu of functioning institutions. In those cases, states only maintain power and authority through oppression and the use of repressive tactics. However, in the end state repression can only subdue the populace for so long in the face of deteriorating conditions due to state failure. Political violence becomes the only alternative for repressed citizens of a failed state, and a perfect example of this phenomenon is provided in William Reno’s study of Sierra Leone (Rotberg, 2003). Reno illustrates that due to the corrupt actions of her leaders, widespread conflict was inevitable in Sierra Leone, and yet the emergence of internal conflict did not immediately threaten the power of those whose actions generated the violence (Rotberg 2003). Instead, those political authorities used their own brand of violence and insecurity, instead of 968 T. Howard

political institutions, to control their citizens for several years. “These techniques included manipulating access to remaining economic opportunities, instigating local tensions and factional divisions, and using disorder as a political instrument to divide potential challenges to their authority” (Rotberg 2003, p.71). In the absence of bureaucratic institutions, political authority in Sierra Leone was based on controlling markets and manipulating access to economic opportunities. The state authority’s capacity to privatize state assets played a key role in its ability to exercise authority and legitimacy. Yet eventually the state authority could only maintain a minuscule degree of strength in the presence of bureaucratic weakness, corruption, and economic scarcity. When the fac¸ade faded, the state authority’s strategy to hang on to their position of power was based on their ability to undermine the security of its citizens. Given the state’s interest in creating a security dilemma, conflict became inevitable. With the ineffective leadership of Stevens and the APC (All Peoples Congress), Sierra Leone was dragged from failure into complete collapse once the RUF seized control of the government through the use of terror and political violence.

The Provision of Public Goods H4: When citizens are provided with tangible public goods, they are less likely to participate in and hold support for political violence.

Once the public good of security has been ensured, the provision of other tangible goods such as an education system, health care system, transportation systems, mail delivery system, and other basic services, are essential responsibilities of the state (Debeil and Klein 2002; Kreijen 2004; Rotberg 2003, 2004; Zartman 1995). When states fail to provide these fundamental services, individuals must seek alternative ways in which to access these public goods (Azam 2001). This may be in the form of resorting to political violence themselves, as a way in which to obtain these goods, or individuals may throw their support behind terrorist organizations and insurgent groups because they believe these groups will succeed where the state has failed.

The Protection of Private Property H5: When citizens are provided with adequate protection of their private prop- erty, they are less likely to participate in and hold support for political violence. Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 According to Boone, tensions born out of territorial disputes and environmental stresses are more likely to result in conflict (2007). Bates makes a similar argument. He finds that violent consequences of state decline become evident when private property becomes insecure (2007). Therefore, in states where the state authority engages in rent seeking and predatory behavior, citizens are less likely to feel that their property is protected. Therefore, they may rely on violent organizations to provide this protection, or they may engage in violent activities in order to protest the state authority’s corrupt behavior in the illegal seizure of private property. At the same time, when individuals feel another group or individual is threatening their property, they may turn to organizations that engage in terror and violence to remove this threat, which is often at the root of ethnic and genocidal conflicts. Failed States and Terrorism in Sub-Saharan Africa 969

Data and Methodology

Data The data for this analysis come primarily from the Afrobarometer Survey, which collects data from several sub-Saharan African countries, with approximately 1,200 respondents of voting age from each state. Based upon a random representative national sample, the survey was conducted in three rounds and data were gathered using face to face interviews. For the purpose of this study, the third round is used (2005–2006), which consists of a total of eighteen cases (n ∼ 1,200 for each case): South Africa, Namibia, Botswana, Lesotho, Ghana, Zimbabwe, Nigeria, Zambia, Tanzania, Uganda, Malawi, Mali, Cape Verde, Kenya, Mozambique, Senegal, Benin, and Madagascar.2 As discussed at the opening of this article, the overwhelming majority of Africa’s nations are considered failed states. An examination of the 2009 Index of State Weakness reveals that with the exception of Cape Verde, South Africa, and Botswana, the remaining fifteen cases in the sample rank in the bottom three quintiles (Table 1). The 2009 Failed State Index reports similar findings, except it ranks all eighteen of the cases from the Afrobarometer’s Survey as either a state in the phases of warning or alert (Table 2). The indexes illustrate that as a unit of analysis, Africa presents a unique opportunity to observe the impact state failure has on popular support for the use of political violence. Furthermore, the sample is representative of this particular phenomenon in the region, which is why all eighteen cases are included in the analysis.

Table 1 The 2009 index of state weakness rankings for the country cases in round three of the Afrobarometer Survey

Country Index of State Weakness Ranking Quintile Zimbabwe 8 Bottom Quintile Uganda 27 Bottom Quintile Nigeria 28 Bottom Quintile Zambia 32 2nd Quintile Mozambique 39 2nd Quintile Malawi 46 2nd Quintile Madagascar 49 2nd Quintile Kenya 50 2nd Quintile

Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 Mali 52 2nd Quintile Lesotho 53 2nd Quintile Tanzania 55 2nd Quintile Senegal 68 3rd Quintile Benin 71 3rd Quintile Namibia 82 3rd Quintile Ghana 84 3rd Quintile Botswana 102 4th Quintile South Africa 110 4th Quintile Cape Verde 123 5th Quintile (The Highest)

Rankings based on a total of 141 Countries. 970 T. Howard

Table 2 The 2009 Failed State Index rankings for the country cases in round three of the Afro- barometer Survey

Country Failed State Index Ranking Quintile Zimbabwe 2 Alert Kenya 14 Alert Nigeria 15 Alert Uganda 21 Alert Malawi 28 Alert Zambia 60 Warning Lesotho 67 Warning Madagascar 68 Warning Tanzania 70 Warning Mozambique 72 Warning Mali 83 Warning Cape Verde 84 Warning Namibia 96 Warning Benin 97 Warning Senegal 102 Warning Botswana 116 Warning South Africa 122 Warning Ghana 124 Warning

Rankings based on a total of 177 countries.

In addition to the Afrobarometer Survey data, this study also relies on data from the Na- tional Consortium for the Study of Terrorism and Response to Terrorism (START)—Global Terrorism Database (GTD). The Global Terrorism Database includes information on global terrorist events for the time period of 1970–2007. With a record of more than 80,000 cases that covers incidents for nearly four decades, and across all regions of the globe, the Global Terrorism Database includes systematic data on both international as well as domestic ter- rorist events. This study utilized the GTD to observe the various trends in terrorist activity within sub-Saharan Africa for the thirty-year time period of 1977–2007.

Model Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 This study is largely based on statistical analysis of the survey data. Therefore, the author conducted cross-sectional analysis using logistical regression in order to determine which dimensions of state failure increase the probability that an individual will support political violence in the selected country cases (Figure 1).

⎛ p (violence) ⎞ Model1 = ln⎜ i ⎟ = β + β Security + β StatePresence + β StateAuthority + β StateLegitimacy + ⎜ − ⎟ 0 1 2 3 4 ⎝ 1 pi (violence) ⎠ β + β + β + β + β + β + β + 5 SPublic Goods 6 Property 7 Education 8Gender 9 PersEcon 10 Ethnic 11Urban e

Figure 1. Logit model support for political violence. Failed States and Terrorism in Sub-Saharan Africa 971

Description of the Variables

Dependent Variable—Political Violence The dependent variable is a measure of political violence. It captures individuals’ attitudes regarding the use of political violence as a legitimate tool to secure tangible goods or concessions from the state. The measure for political violence is based on the following question in the Afrobarometer Survey:

Question (Variable = Violence): Which of the following statements is closest to your view. Choose Statement A or Statement B. A: The use of violence is never justified in [respondent’s country] politics B: in this country, it is sometimes necessary to use violence in support of a just cause. Value Labels: 1 = Agree Very Strongly with A, 2 = Agree with A, 3 = Agree with B, 4 = Agree Very Strongly with B, 5 = Agree with Neither, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data3

In order to estimate the model, the values “1” and “2” were set to zero, values “3” and “4” to one, and the remaining values coded as Missing Data. This leaves 23,993 observations out of an original 25,397 observations. The reasoning for coding the variable in this manner is because the author is only interested in those respondents that favor the use of violence in relation to those that do not. The motivations of respondents who do not have an answer, or agree with neither statement, are not known, so this information is not useful. In the end, the transformation of the variable allows one to determine which dimensions of state failure increase the likelihood that individuals will support the use of political violence. Therefore, the value “1” indicates the respondent agrees with Statement B and supports the use of political violence, whereas the value “0” indicates the respondent agrees with Statement A and does not support the use of political violence.

Explanatory Variables The Public Good of Security. The first explanatory variable discussed is based on two measures of the public good of security. Together, these measures capture individuals’ opinions regarding how secure they feel. The belief is that if an individual’s feelings of personal insecurity are great, the more willing they are to turn to political violence as a way of actively demonstrating against the state that has failed to protect them. They are also more willing to support groups that engage in political violence because they believe these Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 groups will provide protection for them. For example, in 2006, the Beirut Center for Research and Information found in a survey taken in July that approximately 87 percent of the Lebanese citizens polled supported Hezbollah, a known terrorist organization, and their actions against Israel (Beirut Center for Research and Information 2006). The same was true of Palestinian support for Hamas and Hezbollah. Following the capture of an Israeli soldier in 2006, a 24 July survey conducted by An-Najah University in the West Bank town of Nablus found that 91 percent of Palestinians supported the militants’ demands that the soldier only be released in exchange for prisoners held by Israel (Beirut Center for Research and Information 2006). It is widely believed that Hezbollah and Hamas (Wilson 2006, 2009) enjoy the support of the Lebanese and Palestinians largely because these terrorist organizations are perceived as “protectors 972 T. Howard

from Israeli aggression” (Aslan 2006; Jamail 2006). Therefore, the greater one’s personal insecurity, the more likely they are to support the use of political violence. The measures for the public good of security are based on the following questions in the Afrobarometer Survey:

Question 1(Variable = Security1): Over the past year, how often (if ever) have you or anyone in your family: Feared crime in your own home. Value Labels: 0 = Never, 1 = Just once or twice, 2 = Several times, 3 = Many times, 4 = Always, 9 = Don’t Know, 998 = Refused to Answer, –1 = Missing Data Question 2(Variable = Security2): Over the past year, how often (if ever) have you or anyone in your family: Been physically attacked. Value Labels: 0 = Never, 1 = Just once or twice, 2 = Several times, 3 = Many times, 4 = Always, 9 = Don’t Know, 998 = Refused to Answer, –1 = Missing Data

In order to estimate the model, the values of “9” and “998” were coded as Missing Data for both variables. By removing these values, a scale is left that indicates an increase in perceived insecurity. Given the measurement of this explanatory variable, a positive coefficient indicates that a higher level of perceived insecurity increases the likelihood that a respondent will support political violence.

The Presence of the State. The second explanatory variable is the presence of the state, and includes three measures of this dimension. Together, these measures capture individu- als’ opinions regarding the frequency and duration of contact from their elected officials. The belief is that the more visible the government, the more likely the state is present, especially in the case where the survey asks the respondent to indicate how often they have contacted their local and national government in the past year to solve a problem. Individuals are less likely to contact their government officials to solve a problem if they perceive the government’s presence in their daily lives to be negligible. This situation would support the hypothesis regarding the presence of the state and ungoverned territory. A state where the elected officials are perceived as present, active, and visible, the less likely the region these officials oversee would be considered ungoverned territory. The measures for the presence of the state are based on the following questions in the Afrobarometer Survey:

Question 1(Variable = Presence1): During the past year, how often have you Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 contacted any of the following persons for help to solve a problem or to give them your views: A Local Government Councillor? Value Labels: 0 = Never, 1 = Only once, 2 = A few times, 3 = Often, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 2 (Variable = Presence2): During the past year, how often have you contacted any of the following persons for help to solve a problem or to give them your views: A Member of Parliament/National Assembly Representative? Value Labels: 0 = Never, 1 = Only once, 2 = A few times, 3 = Often, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 3 (Variable = Presence3): How much time does your member of Parliament/National Assembly Representative spend in this constituency? Failed States and Terrorism in Sub-Saharan Africa 973

Value Labels: 0 = Never, 1 = At Least Once a Year, 2 = At Least Once A Month, 3 = At Least Weekly, 4 = Here Almost All of the Time, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data

In order to estimate the model, the values of “9” and “98” were coded as Missing Data for the three variables. By removing these values, a scale is left that indicates an increase in the perceived presence of the state for all measures. Given the measurement of these independent variables and the hypothesis, a negative coefficient would indicate that the perceived absence of the state increases the likelihood that a respondent will participate in or support political violence.

The State Authority. The third explanatory variable, state authority, is based on seven measures of this dimension. Together, these measures capture individuals’ opinions regard- ing the perceived corruptness and legitimacy of the government and its elected officials. The belief is that if the state authority is perceived as corrupt, or as having seized power through illegal means, it is difficult for the state to convince citizens to adhere to basic laws and institutional rules. In support of the hypothesis, a corrupt or illegitimate state author- ity encourages citizens to engage in political violence as a means of protest against the government. The measures for the dimensions of state authority that capture government corruption are based on the following questions in the Afrobarometer Survey:

Question 1(Variable = Corrupt1): How many of the following people do you think are involved in corruption, or haven’t you heard enough to say: The President/Prime Minister and Officials in his Office? Value Labels: 0 = None, 1 = Some of them, 2 = Most of them, 3 = Allofthem, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 2 (Variable = Corrupt2): How many of the following people do you think are involved in corruption, or haven’t you heard enough to say: Members of Parliament/National Assembly Representatives? Value Labels: 0 = None, 1 = Some of them, 2 = Most of them, 3 = Allofthem, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 3 (Variable = Corrupt3): How many of the following people do you think are involved in corruption, or haven’t you heard enough to say: Local Government Councillors? Value Labels: 0 = None, 1 = Some of them, 2 = Most of them, 3 = Allofthem, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 4 (Variable = Corrupt4): How many of the following people do you think are involved in corruption, or haven’t you heard enough to say: National Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 Government Officials? Value Labels: 0 = None, 1 = Some of them, 2 = Most of them, 3 = Allofthem, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 5 (Variable = Corrupt5): How many of the following people do you think are involved in corruption, or haven’t you heard enough to say: Local Government Officials? Value Labels: 0 = None, 1 = Some of them, 2 = Most of them, 3 = Allofthem, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data

In order to estimate the model, the values of “9” and “98” were coded as Missing Data for the variables. By removing these values, a scale is left that indicates an increase 974 T. Howard

in the perceived presence of corruption among elected officials for all measures. Given the measurement of these independent variables and the hypothesis, a positive coefficient would indicate that the perceived presence of government corruption increases the likelihood that a respondent will support political violence. Turning now to the other dimensions of state authority, the article employs a set of measures that also capture the legitimacy of the state authority by measuring the process by which officials are elected to office, and if this is a fair and equitable process. The measures for the legitimacy of the state authority are based on the following questions in the Afrobarometer Survey:

Question 1 (Variable = Legitimate1): Think about how elections work in this country. How well do elections ensure that the Members of Parliament/National Assembly Representatives reflect the views of the voters? Value Labels: 0 = Not at all well, 1 = Not very well, 2 = Well, 3 = Very Well, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 2 (Variable = Legitimate2): Think about how elections work in this country. How well do elections enable voters to remove from office leaders who do not do what the people want? Value Labels: 0 = Not at all well, 1 = Not very well, 2 = Well, 3 = Very Well, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data

In order to estimate the model, the values of “9” and “98” were coded as Missing Data for both variables. By removing these values, a scale is left that indicates an increase in the perceived legitimacy of the state. Given the measurement of these independent variables and the hypothesis, a negative coefficient would indicate that the perceived absence of government legitimacy increases the likelihood that a respondent will participate in or support political violence.

The Provision of Public Goods. The fourth explanatory variable, the provision of public goods, is based on three measures of this dimension. Together, these measures capture individuals’ opinions regarding the state’s ability to provide citizens with essential public goods. The belief is that if the state cannot provide essential public goods, individuals begin to perceive that state as weak or impotent, and turn to other means of obtaining these essential goods. Resorting to crime, corruption, and the black market economy are just a few ways in which individuals gain access to these goods; political violence represents another such avenue. In support of the hypothesis, a state that does not provide essential public goods to its constituency encourages citizens to engage in, or support, political violence Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 as a means of gaining access to these necessary goods. The measures for the provision of public goods are based on the following questions in the Afrobarometer Survey:

Question 1(Variable = Goods1): Over the past year, how often, if ever, have you or anyone in your family gone without: Enough clean water for home use? Value Labels: 0 = Never, 1 = Just once or twice, 2 = Several times, 3 = Many times, 4 = Always, 9 = Don’t Know, 998 = Refused to Answer, –1 = Missing Data Question 2 (Variable = Goods2): Based on your experience, how easy or difficult is it to obtain the following services? Or do you never try to get these services from the government: A place in primary school for a child? Failed States and Terrorism in Sub-Saharan Africa 975

Value Labels: 1 = Very difficult, 2 = Difficult, 3 = Easy, 4 = Very easy, 7 = Never try, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Question 3 (Variable = Goods3): Based on your experience, how easy or difficult is it to obtain the following services? Or do you never try to get these services from the government: Medical treatment at a nearby clinic? Value Labels: 1 = Very difficult, 2 = Difficult, 3 = Easy, 4 = Very easy, 7 = Never try, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data

For the variable Goods1, the values of “9” and “998” were coded as Missing Data. By removing these values, a scale is left that indicates the increasing ability of the government to provide the public good of clean water. Given the measurement of this variable and the hypothesis, a negative coefficient would indicate that the perceived inability of the government to provide the public good of clean water increases the likelihood that a respondent will support political violence. For the variables Goods2 and Goods3, the values of “9” and “98” were coded as Missing Data. And instead of dropping value 7, which indicates citizens do not try to obtain certain basic services from the government, it was recoded as “0.” This was done because the author does not know the motivations behind an individual’s reasons for not attempting to obtain these particular services from the government. In some cases, wealthy individuals do not need to obtain certain basic services, such as education and health care from the government. Therefore, while the government may provide the service, they simply do not use it. Whereas, individuals living in poverty may need these services, but could be apathetic toward the government and its ability to provide them with these services. Since the author does not have any insight regarding why these respondents never seek to obtain these services from the government, this value cannot be dropped. Important information would be dropped if this was done, which is why the response was coded as zero. After recoding this variable, a scale is left that indicates the increasing ability of the government to provide the public goods of education and health care. Given the measurement of this variable and the hypothesis, a negative coefficient would indicate that the perceived inability of the government to provide the public goods of health care and education increases the likelihood that a respondent will support political violence.

The Protection of Private Property. The fifth explanatory variable is based on a single measure of the protection of private property. This measure captures individuals’ opinions regarding the security of their private property. The belief is that if an individual’s feelings of property insecurity are great, the more willing they are to turn to political violence as a way of actively demonstrating against the state that has either taken their property through Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 illegal means, or allowed their property to be illegally seized by another citizen. Similar to the actions of individuals who lack personal security, and their support for groups who engage in political violence, individuals are also more willing to support groups that engage in political violence because they believe these groups will provide property protection for them, or help them regain access to their property (Bates 2007). The measure for the protection of private property is based upon the following question in the Afrobarometer Survey:

Question 1(Variable = Property): What if anything, would you do to try and resolve each of the following situations: Someone wrongly seized your family’s land? 976 T. Howard

Value Labels: 1 = Don’t worry, things will be resolved given enough time, 2 = Lodge a complaint through proper channels or procedures, 3 = Use connections with influential people, 4 = Offer tip or bribe, 5 = Join in public protest, 6 = Other, 7 = Nothing, because nothing can be done, 9 = Don’t Know, 998 = Refused to Answer, –1 = Missing Data

In order to estimate the model, the values of “9” and “98” were coded as Missing Data for both variables. The author also dropped the value “6” and recoded value “7” as zero.4 In the case of value “6,” which represents “Other,” this was removed because it offers no substantive information for this analysis. In the case of value “7,” which represents “Nothing, because nothing can be done,” the author chose to set this value to zero, because as far as tangible meanings regarding the scale, it is similar to value “1,” which represents “Don’t worry, things will be resolved given enough time.” Both options indicate that nothing is done, the only difference is the motivation behind why the respondent chooses not to

Table 3 List of variable names and description of the concepts measured

Variable name Indicator concept Violence Measures political violence Protest Measures participation in protest or demonstration Security1 Fear of crime Security2 Personally attacked Presence1 Contacts local government to solve problems Presence2 Contacts national government to solve problems Presence3 Time the national government leaders spend with the constituency Corrupt1 The president/prime minister is corrupt Corrupt2 The National Assembly/Members of the Parliament are corrupt Corrupt3 The local councillor is corrupt Corrupt4 National government officials are corrupt Corrupt5 Local government officials are corrupt Legitimate1 Members of the national government govern according to the mandate of the people Legitimate2 The ease with which citizens can remove ineffective leaders Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 from office Goods1 Gone without clean water Goods2 Government provides education Goods3 Government provides medical treatment Property Illegal seizure of private property Education Respondent’s highest level of education Gender Gender PersEcon Respondent’s current economic level Ethnic The respondent’s ethnic group and their current level of economic status in relation to other ethnic groups Urban Respondent is from an urban or rural area Failed States and Terrorism in Sub-Saharan Africa 977

act. By setting value “7” to zero, a scale is left that moves along the continuum from doing nothing out of apathy, to actively protesting. The transformation of this variable indicates an increase in the support for more radical actions, given the loss of private property. Given the measurement of this variable and the hypothesis, a positive coefficient would indicate the more radical the action a respondent is willing to take if their private property has been seized increases the likelihood that a respondent will participate in or support political violence. Controls. The study operationalizes a set of control variables that capture the socioe- conomic conditions of the respondents. One of the most salient differences between the respondents in this study is their socioeconomic level. Otherwise stated, support for political violence is more likely among the “have nots” and the marginalized populations of society. Therefore, the variables, level of education, personal economic situation, ethnic groups’ economic situation, gender, and whether the respondent lives in an urban or rural area are included. These control variables serve as ideal indicators of an individual’s socioeconomic position within that country. See Appendix A for a description of the control variables and their measurement.

Findings

Trend Analysis of State Failure Events and Terrorist Incidents in Sub-Saharan Africa Turning now to the first set of findings, using the 2007 START Data, the article reports the trend of terrorist activity in sub-Saharan Africa for the thirty-year time period of 1977–2007 (Figure 2). Figure 2 reveals that in the seventies and early eighties, incidents of terrorism in the region were relatively low, with attacks averaging around fifty per year. However, toward the end of the eighties and into the nineties, there was a sharp increase in terrorism. Arguably, this was due to the failure of the post-independence governments to promote stability, comprehensive democracy, and economic development. With the majority of African independence movements taking place in the sixties and seventies (see Appendix C), there was about a fifteen-year period in the region where leaders struggled to implement democracy and promote stability. However, when these attempts failed, rebel groups emerged and civil war erupted, plunging many of sub-Saharan Africa’s states into failure. As a consequence, political violence erupted during these periods of failure, one of which that occurred during the tumultuous period of the nineties. Another era where one can observe a spike in political violence has begun in the latter part of the twenty-first century. This is largely the result of the failure of post–civil war peace agreements. Again, the people of Africa have been disappointed by the corruption and polarizing factions of the political leadership, thus opening the door for internationally sponsored terrorist organizations to offer an alternative strategy for obtaining political and Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 economic goods that the government has failed to provide. The major problem with the current surge in political violence and the growing presence of terrorist groups in the region is that unlike the period before in the nineties, radical Islamic ideologies have found a home in sub-Saharan Africa and have begun to take root, making the presence of these groups, coupled with the failure of African states, a very deadly combination that the international community must now contend with. Statistical Results Turning now to the statistical results, the findings echo the trends observed in Figure 2. The conditions of state failure foster individual support for the use of political violence. (See Table 3 for a list of variable names and descriptions of the concepts measured). Table 4 978 T. Howard

Table 4 Binary logit estimates of support for political violence Variable Restricted model Full model Security1 –.004 –.0045 (.021) (.022) Security2 .171∗∗∗ .175∗∗∗ (.035) (.036) Presence1 –.004 –.008 (.027) (.028) Presence2 .029 .035 (.041) (.041) Presence3 .114∗∗∗ .093∗∗∗ (.021) (.022) Corrupt1 .165∗∗∗ .185∗∗∗ (.038) (.040) Corrupt2 –.028 –.040 (.045) (.046) Corrupt3 –.005 –.029 (.041) (.042) Corrupt4 –.072 –.060 (.046) (.047) Corrupt5 .060 .056 (.043) (.044) Legitimate1 –.030 –.040 (.032) (.033) Legitimate2 –.072∗∗ –.064∗∗ (.029) (.030) Goods1 .009 .024 (.018) (.019) Goods2 –.049∗∗ –.049∗∗ (.022) (.023) Goods3 –.033 –.054∗∗ (.026) (.027) Property .132∗∗∗ .129∗∗∗ (.031) (.032) Gender — .040 (.049) PersEcon — .056∗∗ (.021) Education — .048∗∗∗ Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 (.013) Ethnic — –.07∗∗ (.025) Urban — .057 (.053) Constant –1.61∗∗∗ –1.67∗∗∗ (.120) (.173) Observations 11380 10789 Pseudo R2 .0139 .0178 LR χ 2 159.37 194.08 Prob >χ2 .000 .000

∗∗∗p < .01 for two-tailed test; ∗∗p < .05 for two-tailed test; standard errors in parentheses. Failed States and Terrorism in Sub-Saharan Africa 979

Figure 2. Number of reported terrorist incidents in sub-Saharan Africa, 1977–2007. Source: “Global Terrorism Datatbase.” START. http://www.start.umd.edu/gtd/ (accessed 23 January 2010). Data com- piled by author.

reports the binary logit estimates for the model which indicates the probability that a respondent will support the use of political violence. Table 5 reports the marginal effects for the logit model. Both are interpreted in this discussion. (See Appendix B for descriptive statistics of the variables).

Findings that Are Statistically Significant and Support the Hypotheses. Regarding the statistical findings, the first set of variables discussed are those that are both statistically significant and provide support for the stated hypotheses. The first of such variables is Security2. This variable indicates whether the respondent, or someone in the respondent’s family, has been personally attacked in the past year. The findings suggest that a higher level of perceived physical insecurity increases the probability that a person will support the use Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 of political violence. A one standard deviation increase in a person’s personal insecurity leads to an approximately 3 percent increase in support for political violence as a legitimate political tool. In terms of the concept of state authority, the variable Corrupt1 indicates whether the respondent perceives the national government as being corrupt. The findings suggest that a higher level of perceived corruption within the national government increases the probability that a person will support the use of political violence. A one standard deviation increase in a person’s belief that the national government is corrupt leads to an approximately 3 percent increase in support for political violence as a legitimate political tool. The second measure of state authority is legitimacy, and the findings suggest that the more difficult it is for voters to remove an ineffective leader from office, the more supportive 980 T. Howard

Table 5 Marginal change in variables of significance for Model 1-support for political violence

Variable Marginal change Percentage change Support for hypotheses Security2 .028 2.8Yes Presence3 .015 1.5No Corrupt1 .029 2.9Yes Legitimate2 −.010 −1Yes Goods2 −.008 −0.8Yes Goods3 −.009 −0.9Yes Property .021 2.1Yes PersEcon .009 0.9No Education .008 0.8Yes∗ Ethnic −.012 −1.2No

The marginal effect indicates an increase in the predicted probability of the dependent variable according to one standard deviation change in the independent variable, holding all other variables constant. ∗In the case of this analysis, education can be interpreted in either direction as a variable concept that supports the hypothesis.

they are of using political violence as an option. A one standard deviation increase in a person’s ease with which they can remove an ineffective leader from office leads to a 1 percent decrease in support for political violence as a legitimate political tool. Turning now to the provision of public goods, the findings suggest that the more difficult it is for respondents to receive the public goods of education and health care from the government, the more supportive they are of using political violence as an option. A one standard deviation increase in a person’s ease with which they can obtain the public good of education leads to a .8 percent decrease in support for political violence as a legitimate political tool. And a one standard deviation increase in a person’s ease with which they can obtain the public good of health care leads to a .9 percent decrease in support for political violence as a legitimate political tool. As far as the protection of private property is concerned, the more radical the action a respondent is willing to take if their private property has been seized increases the likelihood that a respondent will support the use of political violence. A one standard deviation increase in a person’s willingness to use radical action if their personal property has been unlawfully seized leads to an approximately 2 percent increase in support for political violence as a legitimate political tool. Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 With regard to the control variable, Education, it can be interpreted either way. The prevailing belief is that low levels of education are associated with poor living standards and poor economic conditions (Smith 2005; Sen 1982; Brady 2003; Danziger and Wein- berg 1994; Berliner 2005; Becker and Luthar 2002; Tilak 2002; Orfield and Lee 2005). As a consequence, this group would be more willing to support political violence as a viable political option in order to increase their economic status and improve their living conditions. At the same time, the country cases included in Round 3 of the Afrobarometer Survey are those that would be considered liberalizing regimes. Therefore, as Gurr (1970) and Huntington (1968) demonstrate, countries that are modernizing increase expectations among the middle class, who are typically educated; however, the economic achievements of an evolving economy are not always in line with citizens’ expectations. Consequently, Failed States and Terrorism in Sub-Saharan Africa 981

individuals who are educated may support the use of political violence because they believe the government is not delivering on its promise to modernize. The results of this analysis would support this proposition, in that the study finds that the more educated the respondent, the more likely they are to support the use of political violence. A one standard deviation increase in a person’s level of education leads to an approximately .8 percent increase in support for political violence as a legitimate political tool.

Findings that Are Statistically Significant but Fail to Support the Hypotheses. With regard to the results, the second set of variables discussed are those that are statistically significant but fail to provide support for the hypotheses. The first of such variables is Presence3. This variable indicates the amount of time the respondent’s Member of Parliament or National Assembly Representative spends with his or her constituency. The findings suggest that the perceived presence of the state increases the probability that a person will support the use of political violence, which contradicts the hypothesis. As far as personal living conditions are concerned, the results indicate that the better the respondent’s living conditions, the more likely they are to support the use of political violence. Unlike the variable, Education, which is naturally highly correlated with an indi- vidual’s living and economic conditions,5 there is little justification to interpret this variable in support of the hypothesis. That is because the level of education for an individual confers on that person a certain ideology and expectation that with higher levels of education, comes a higher economic status, and improved living conditions. That is the expectation, but often that is not the case in these particular nations. However, a person’s economic status and living conditions are tangible indicators that the government has delivered on

Table 6 Hypotheses supported by the logistical regression analysis

Hypotheses Variable and concept measured Absence of the public good of security Security2—Personally attacked State authority Corrupt1—The president/Prime minister is corrupt Corrupt3—The local councillor is corrupt Corrupt5—The local government is corrupt State legitimacy Legitimate2—The ease with which

Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 citizens can remove ineffective leaders Provision of public goods Goods2—Government provides education Goods3—Government provides healthcare Protection of private property Property—Illegal seizure of private property Controls Gender Education—Respondent’s highest level of education 982 T. Howard

its promises. At the least, higher economic status and satisfactory living conditions should indicate satisfaction with the government. Therefore, one would not expect a respondent who is living well and comfortably to support the use of political violence. The final variable to achieve statistical significance, but contradicts the hypothesis is the measure of ethnicity. This variable indicates whether a respondent feels their ethnic group is doing worse, the same, or better than other ethnic groups with regards to their current living conditions. The findings suggest that a respondent’s belief that their ethnic group is doing better than other groups increases the probability that they will support the use of political violence. One would expect only those respondents who feel their ethnic group is worse off than other groups to support the use of political violence. Therefore, it is apparent that this variable fails to support the hypothesis.

Conclusion This study has argued that state failure increases the likelihood that citizens will support the use of political violence in these fragile sub-Saharan African states. Out of the five dimensions of state failure, four have a significant impact on the likelihood that a citizen will support the use of political violence (see Table 6). The findings point to the absence of the public good of security, a corrupt and illegitimate state authority, lack of provision of essential public goods, and the inability to protect private property as the major components of state failure that increase the probability of political violence. The unexpected exception to this is the failure of this analysis to find support for the assertion that with ungoverned territory there is the absence of the state. The belief is that without the presence of the government, there is the potential for non-state actors to organize themselves in these spaces and engage in illegal and dangerous activities without having to contend with the interference of the central government. These non-state actors can be citizens who have organized themselves into militia organizations,6 or non-citizens who have crossed into that state.7 Yet, regardless of where these groups come from, and who they are comprised of, the main argument remains that the absence of the government is a significant precursor to political violence. And yet, this analysis fails to find evidence to support this argument, which represents an important avenue for future research. As the author looks to continue this vein of research, she is focused on finding more accurate measures of the absence of the state. The author remains grounded in the notion that this feature of state failure leads to political violence, and intend to work toward incorporating more precise measures of this concept. At the same time, the author also looks to extending this research to other regions, and utilizing additional data sources, such as the Latin American Public Opinion Project, Latinobarometer, and the World Values Survey data to estimate similar models of citizen’s support for and participation in political Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 violence. While, sub-Saharan Africa remains unique in that the dimensions of state failure can be more readily observed in this region, it is anticipated that the model estimated in this analysis will achieve similar results for Latin America and the Middle East because the dimensions of state failure that drive public support for political violence remain the same regardless of the region. To conclude, this study demonstrates that in sub-Saharan Africa where state failure is a tragic pandemic, there is evidence to suggest that individuals are likely to support terrorism and political violence as a means by which to obtain tangible political, economic, and social goods. This study has important political implications because thus far, sub-Saharan Africa has not produced the level of terrorist activity that is present in other regions, such as the Middle East and South Asia. However, that may not always be the case. As this Failed States and Terrorism in Sub-Saharan Africa 983

article has demonstrated, the conditions in failed states threaten an individual’s survival, which then encourages them to support the use of terrorism and political violence to obtain tangible political and economic resources. Therefore, this phenomenon raises the question could the failed states of sub-Saharan Africa become the next breeding ground for internationally sponsored networks of terrorism? Undoubtedly, yes, as it has already begun, but for how long will these networks find a safe haven in the region, and how deadly will this terrorist threat be to both the region and the global community? Only time can provide that answer.

Notes

1. Resistance movements are considered nonviolent and represent those groups that engage in civil disobedience. When they begin to use violence, they transform into rebel or insurgent groups. 2. It is important to note that the Afrobarometer Survey data are intentionally biased toward liberalizing regimes; therefore, authoritarian regimes and countries embroiled in conflict are under- represented. 3. All missing data, including the value label “–1 = Missing Data” are removed from the analysis. 4. By dropping a value, it is coded as missing data. 5. Pearson correlation coefficient = .2018, p = .000. 6. Examples of this include the Sudan People’s Liberation Movement/Army (SPLA/M) led by John Garang, and comprised of the Dinka and Nuer people from the South and the Sudan Liberation Movement/Army (SLM/A), which is comprised of three non-Arab groups—the Masalit, Fur, and Zaghawa. 7. Al Qaeda, Hezbollah, the Palestinian Islamic Jihad, Hamas, the Egyptian Islamic Group, and the Abu Nidal organization have all at one point crossed into Sudan to set up operations (Linden 2002).

References

Abedin, Mahan. 1998. Asymmetric threats. In Strategic Assessment 1998. Washington, DC: Institute for National Security Studies. Adamson, C. 1998. Tribute, turf, honor, and the American street gang: Patterns of continuity and change since 1820. Theoretical Criminology 2(1): 57–84. Addison, Michael. 2002. Violent politics: Strategies of internal conflict. Basingstoke: Palgrave. Akers, R. L. 1998. Social learning and social structure: A general theory of crime and deviance. Boston, MA: Northwestern University Press. Akers, R. L., and L. Gang. 1999. Age, social learning, and social bonding in adolescent substance Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 use. Deviant Behavior, 19: 1–25. Aslan, Reza. 2006. Hezbollah is nobody’s puppet. Los Angeles Times. 25 July. Azam, Jean-Paul. 2001. The redistributive state and conflicts in Africa. The Journal of Peace Research 38(4): 429–444. Bates, Robert. 1981. Markets and states tropical Africa: The agricultural basis of agricultural policies. Berkeley and Los Angeles: University of California Press. Bates, Robert. 2007. Political conflict and state failure. In The political economy of economic growth in Africa, 1960–2000, eds. Benno Ndulu, Stephen O’Connell, Robert Bates, Paul Collier, and Charles Soludo, 249–296. Cambridge: Cambridge University Press. Becker, Bronwyn E. and Suniya S. Luthar. 2002. Social-emotional factors affecting achievement out- comes among disadvantaged students: Closing the achievement gap. Educational Psychologist 37(4): 197–214. 984 T. Howard

Berliner, David C. 2005. Our impoverished view of educational reform. Teachers College Record. 2 August. Beirut Center for Research and Information. 2006. Poll finds overwhelming majorities in Lebanon support Hezbollah, distrust U.S. World Public Opinion: Global Public Opinion International Affairs. 2 August. Available at http://www.worldpublicopinion.org Boone, Catherine. 2007. Property and constitutional order: Land tenure reform and the future of the African state. African Affairs (October): 1–29. Brady, David. 2003. Rethinking the sociological measurement of poverty. Social Forces 81(3): 715–751. Chandler, David, ed. 2009. Statebuilding and intervention: Policies, practices and paradigms. Rout- ledge Studies in Intervention and Statebuilding. London: Routledge. Chesterman, Simon, Michael Ignatieff, and Ramesh Thakur. 2005. Making states work: State failure and the crisis of governance. Tokyo: The United Nations University Press. Chomsky, Noam. 1988. The culture of terrorism. Boston: South End Press. Cilluffo, Frank, ed. 2001. Combating chemical, biological, radiological, and nuclear terrorism: A comprehensive strategy: A report of the CSIS Homeland Defense Project. Washington, DC: Center for Strategic and International Studies. Cooley, John. 2000. Unholy wars: Afghanistan, America and international terrorism. London: Pluto Press. Craig, Captain D. W. 1988. Asymmetrical warfare and the transnational threat: Relearning the lessons from Vietnam. Ontario: Canadian Forces College. Danziger, Sheldon H. and Daniel H. Weinberg. 1994. The historical record: Trends in family in- come, inequality, and poverty. In Confronting poverty: Prescriptions for change, eds. Sheldon H. Danziger, Gary D. Sandefur, and Daniel H. Weinberg, 18–50. New York: Russell Sage Foundation. Debiel, Tobias and Axel Klein. 2002. Fragile peace: State failure, violence and development in crisis regions. London: Zed Books Ltd. Elliot, D. D. and S. Menard. 1996. Delinquent friends and delinquent behavior: Temporal and developmental patterns. In Delinquency and crime: Current theories, ed. D. Hawkins, 28–67. Cambridge: Cambridge University Press. Esty, Daniel C., Jack A. Goldstone, Ted Robert Gurr, Barbara Harff, Marc Levy, Geoffrey D. Dabelko, Pamela T. Surko, and Alan N. Unger. 1995. State Failure Task Force Report: Phase I Findings. McLean, VA: Science Applications International Corporation. Esty, Daniel C., Jack A. Goldstone, Ted Robert Gurr, Barbara Harff, Marc Levy, Geoffrey D. Dabelko, Pamela T. Surko, and Alan N. Unger. 1998. State Failure Task Force report: Phase II findings. McLean, VA: Science Applications International Corporation. 31 July. Flannery, Daniel J. and Ronald C. Huff, eds. 1999. Youth violence: Prevention, intervention, and social policy. Washington, DC: American Psychiatric Press. Franks, Jason. 2006. Rethinking the roots of terrorism. New York: Palgrave Macmillan. Freedman, Lawrence, ed. 2002. Superterrorism: Policy responses. Oxford, UK: Blackwell Publishing. Goldstone, Jack, Ted Robert Gurr, Barbara Harff, Marc A. Levy, Monty G. Marshall, Robert H. Bates, Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 David L. Epstein, Colin H. Kahl, Pamela T. Surko, John C. Ulfelder, and Alan N. Unger. In consultation with Matthew Christenson, Geoffrey D. Dabelko, Daniel C. Esty, and Thomas M. Parris. State Failure Task Force report: Phase III findings. McLean, VA: Science Applications International Corporation. Gurr, Ted Robert. 1970. Why men rebel. Princeton, NJ: Princeton University Press. Harbeson, John W. and Donald Rothchild, eds. 2008. Africa in world politics: Reforming political order, 4th edition. Boulder, CO: Westview Press. Hoffman, Bruce. 2006. Inside terrorism. New York: Columbia University Press. Howard, Tiffiany. 2008. Revisiting state failure: Developing a causal model of state failure based upon theoretical insight. Civil Wars 10(2): 125–296. Howard, Tiffiany. 2010. The tragedy of failure: Evaluating state failure and its impact on the spread of refugees, terrorism, and war. Santa Barbara, CA: Praeger Security International-Greenwood Publishing. Failed States and Terrorism in Sub-Saharan Africa 985

Huntington, Samuel. 1968. Political order in changing societies. Ithaca, NY: Yale University Press. Huntington, Samuel. 1991. The third wave: Democratization in the late twentieth century. London: Oklahoma University Press. Jackson, Robert H. and James Alan. 1993. States in a changing world. Oxford: Carendon Press. Jackson, R. H. and C. G. Rosberg. 1982. Personal rule in black Africa. Berkeley, CA: University of Berkeley Press. Jamail, Dahr. 2006. Hezbollah could be gaining strength. Inter Press News Service. 26 July. Klein, Malcom and Cheryl Maxson. 2006. Street gang patterns and policies. Oxford: Oxford Uni- versity Press. Kreijen, Gerard. 2004. State failure, sovereignty and effectiveness. Leiden, The Netherlands: Martinus Nijhoff Publishers. Levi, Margaret. 1988. Of Rule and Revenue. Berkeley, CA: University of California Press. Linden, Edward. 2002. World terrorism. New York: Nova Science. Long, David. 1990. The anatomy of terrorism. New York: Free Press. Martin, R. C. 1987. Religious violence in Islam: Towards an understanding of the discourse on Jihad in modern Egypt. In Contemporary research on terrorism, ed. P. Wilkinson and A. Stewarts, 55–72. Aberdeen, U.K.: Aberdeen University Press. Menkhaus, Ken. 2002. Somalia: Next up in the War on Terrorism, CSIS Africa Notes, No. 6, January. Available at www.csis.org/component/option,com csis pubs/task,view/id,2139/type,3/ (accessed 12 February 2009). Miliken, Jennifer. 2003. State failure, collapse and reconstruction. Hoboken, NJ: Blackwell Publishing. Mohammed, Ibrahim. 2009. Somali police warn of more suicide bombings. Reuters. 7 December. Available at http://www.reuters.com/article/idUSTRE5B63DV20091207 (accessed 10 Decem- ber 2009). Newman, Edward. 2007. Weak states, state failure and terrroism. Terrorism and Political Violence 19(4): 463–488. Orfield, G. and C. Lee. 2005. Why segregation matters: Poverty and educational inequality.Cam- bridge, MA: Harvard University and The Civil Rights Project. Piazza, James A. 2006. Rooted in poverty?: Terrorism, poor economic development and social cleavages. Terrorism and Political Violence 18: 159–177. Piazza, James A. 2007. Draining the swamp: Democracy promotion, state failure, and terrorism in 19 Middle Eastern countries. Studies in Conflict and Terrorism 30: 521–539. Rabasa, Angel, Steven Boraz, Peter Chalk, Kim Cragin, Theofore W. Karasik, Jennifer D.P. Mo- roney, Kevin A. O’Brien, and John E. Peters. 2007. Ungoverned territories: Understanding and reducing terrorism risks. Santa Monica, CA: RAND Corporation. Rapoport, David. 1999. Terrorism. In Encyclopedia of violence, peace, and conflict, volume 3, 497–510. San Diego: Academic Publishers. Reno, William. 2003. “Warfare in a Post-State Society.” In State Failure and State Weakness in a Time of Terror, edited by Robert Rotberg, 71–100. Washington, D.C.: Brookings Institution Press. Rotberg, Robert, ed. 2003. State failure and state weakness in a time of terror. Cambridge, MA: Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 World Peace Foundation; Washington, DC: Brookings Institution Press. Rotberg, Robert, ed. 2004. When states fail: Causes and consequences. Cambridge, MA: World Peace Foundation; Washington, DC: Brookings Institution Press. Seifert, Kathryn. 2006. How children become violent: Keeping your kids out of gangs, terrorist organizations, and cults. Boston, MA: Acanthus Publishing. Sen, Amartya. 1982. Choice, Welfare, and Measurements. Oxford: Basil Blackwell. Smith, Stephen, C. 2005. Ending global poverty: A guide to what works.NewYork:Palgrave Macmillan. Snowdon, Ben and David Johnson. 2007. Primer on the Embassy bombings and the U.S. strikes on Su- dan and Afghanistan. Available at http://www.infoplease.com/spot/newsfacts-sudanstrikes.html (accessed 10 December 2009). 986 T. Howard

Sorli, Miriam, Nils Peter Gleditsch, and Havard Strand. 2005. Why is there so much conflict in the Middle East? Journal of Conflict Resolution 49(1): 141–165. Tilak, Jandhyala B. G. 2002. Education and poverty. Journal of Human Development 3(2), 191–207. Tikusis, P. 2009. On the relationship between weak states and terrorism. Behavioral Sciences of Terrorism and Political Aggression 1(1): 66–79. Widner, Jennifer A. 1995. States and statelessness in late twentieth-century Africa. Daedalus 124(3): 129–153. Wilkinson, Paul. 2000. The strategic implications of terrorism. In Terrorism and political violence: A sourcebook, ed. M. L. Sondhi, 19–49. New Delhi, India: Har-anand Publications. Wilson, Scott. 2006. Hamas sweeps Palestinian elections, complicating peace efforts in Mideast. The Washington Post Foreign Service. 27 January. Wilson, Scott. 2009. Palestinian support for Hamas has not declined. The Jewish Institute for National Security Affairs Report, No. 850. 27 January. Zanini, Michele. 1999. Middle Eastern terrorism and netwar. Studies in Conflict and Terrorism 22(3): 247–256. Zartman, I. W. 1995. Collapsed states: The disintegration and restoration of legitimate authority. Boulder, CO: L. Rienner Publishers. Appendix A Control variables 1) Education Question: What is the highest level of education you have completed? Variable Label: Education of respondent Values: 0–9, 98–99, –1 Value Labels: 0 = No formal schooling, 1 = Informal schooling (including Koranic schooling), 2 = Some primary schooling, 3 = Primary school completed, 4 = Some secondary school/ High school, 5 = Secondary school completed/High school, 6 = Post- secondary qualifications, other than university e.g. a diploma or degree from a techni- cal/polytechnic/college, 7 = Some university, 8 = University completed, 9 = Post-graduate, 98 = Refused to Answer, 99 = Don’t Know, –1 = Missing Data ∗In order to estimate the model, the values of “98” and “99” are coded as Missing Data 2) Personal Economic/Living Conditions Question: In general, how would you describe: Your own present living conditions? Variable Label: Your present living conditions Values: 1–5, 9, 998, –1 Value Labels: 1 = Very bad, 2 = Fairly bad, 3 = Neither good nor bad, 4 = Fairly good, 5 = Very good, 9 = Don’t Know, 998 = Refused to Answer, –1 = Missing Data ∗ Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 In order to estimate the model, the values of “9” and “98” are coded as Missing Data 3) Gender Question Number: Q101 Question: Respondent’s gender Variable Label: Gender of respondent Values: 1, 2 Value Labels: 1 = Male, 2 = Female Note: Answered by interviewer. ∗In order to estimate the model, the values are labeled so that “0” equals male and “1” equals female Failed States and Terrorism in Sub-Saharan Africa 987

4) Ethnic group’s economic status Question: Think about the condition of [respondent’s identity group] Are their economic conditions worse, the same as, or better than other groups in this country? Variable Label: Ethnic group’s economic conditions Values: 1–5, 7, 9, 98, –1 Value Labels: 1 = Much Better, 2 = Better, 3 = Same, 4 = Worse, 5 = Much Worse, 7 = Not Applicable, 9 = Don’t Know, 98 = Refused to Answer, –1 = Missing Data Note: Interviewer probed for strength of opinion. If respondent had not identified a group on question 79, this question was marked as “Not Applicable.” ∗In order to estimate the model, the values of “7,” “9,” and “98” are coded as Missing Data 5) Respondent is from an urban or rural area Question: Do you come from a rural or urban area? Variable Label: Interviewer urban or rural Values: 1, 2 Value Labels: 1 = Rural, 2 = Urban Note: Answered by interviewer. ∗In order to estimate the model, the values are labeled so that “0” equals rural and “1” equals urban

Appendix B Descriptive statistics

Variable Observations Mean Std. dev. Min Max violence 23993 .1896803 .3920563 0 1 protest 24692 .1395594 .3465363 0 1 security1 24290 .7142445 1.154027 0 4 security2 24310 .2035788 .6049248 0 4 presence1 24207 .4823398 .907163 0 3 presence2 24206 .1816905 .5812226 0 3 presence3 20736 1.041667 1.104477 0 4 corrupt1 19685 1.173787 .9253391 0 3 corrupt2 19960 1.248697 .8737496 0 3 corrupt3 20562 1.278134 .9010785 0 3 corrupt4 20271 1.352721 .8658986 0 3 corrupt5 20280 1.349162 .8793982 0 3 Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011 legitimate1 22219 1.530492 .9100338 0 3 legitimate2 22187 1.541308 .9762087 0 3 goods1 25347 1.134296 1.381171 0 4 goods2 24568 2.563538 1.256931 0 4 goods3 25048 2.45297 1.045318 0 4 property 23654 2.040374 .6894866 0 5 gender 25397 .4998622 .5000098 0 1 persecon 25308 2.606211 1.189938 1 5 education 25305 3.117803 1.995429 0 9 ethnic 22000 3.161864 .9888373 1 5 urban 25397 .618262 .4858224 0 1 988 T. Howard

Appendix C Map of African Independence, featuring the dates of independence of each nation. Source: http://www.empathosnationenterprises.com/Consulate/EN-Library/Black-Studies/ afindep.html Downloaded by [University of Nevada Las Vegas] at 19:57 03 September 2011